Professional Documents
Culture Documents
May 1998
FINAL REPORT
V
Q
P V
May 1998
Prepared by
____________________________________________________________________________________
* Abbas Abed (SDG&E, Chairman), Joaquin Aquilar (EPE), Nick Chopra (BCH),
Peter Krzykos (APS), Andy Law (WWP), Brian Lee (BCH), Frank McElvain (TSGT),
Saif Mogri (LADWP), Les Pereira (NCPA), Craig Quist (NPC), Ronald Schellberg (IPC),
Joe Seabrook (PSE), Chifong Thomas (PG&E), Boris Tumarin (EPE)
ACKNOWLEDGMENTS................................................................................ iv
2. RECOMMENDATIONS .......................................................................... 5
3. INTRODUCTION.................................................................................. 15
5. DISCUSSION ....................................................................................... 18
7. V-Q ANALYSIS..................................................................................... 33
10.1 Guidelines................................................................................. 67
10.2 Methodologies .......................................................................... 68
APPENDICES
APPENDIX F Excerpts From BPA Blue Ribbon Panel Report ........... 131
LIST OF FIGURES
LIST OF TABLES
On July 2, 1996, and August 10, 1996, two major disturbances occurred within
the Western interconnected region. Both of the disturbances resulted in
significant loss of load and generation throughout the region. Each of the
disturbances was initiated by different events in different locations within the
Northwest area of the Western Interconnection. Two disturbance reports were
generated due to these disturbances. The first report was entitled, WSCC
Disturbance Report for the Power System Outages that Occurred on the Western
Interconnection on July 2, 1996 and July 3, 1996. The second report was
entitled, WSCC Disturbance Report for the Power System Outages that
Occurred on the Western Interconnection on August 10, 1996.
Recommendations from the two reports identified several reliability issues which
were to be investigated. Two issues identified in the disturbance reports pertinent
to this report are reactive power margin studies and undervoltage load shedding
schemes. The Technical Studies Subcommittee (TSS) met on
November 8, 1996 and formed the Reactive Power Reserve Work Group
(RRWG) to address reactive power margin issues.
The RRWG identified the following issues as critical to addressing the reactive
power reserve topic:
The main purpose of this report is to address technical issues regarding reactive
power margins and voltage stability criteria for all member systems. Other
important and related issues, such as equity and financial responsibility, are
beyond the scope of the RRWGs work at this time. These issues can be dealt
with by the RRWG or other appropriate committees after the technical review has
been completed. In this report member system refers to the WECC member
Reactive power margins and voltage stability criteria cover the period after the
transient oscillations have damped out to before the operator can take manual
actions and before area interchange schedules can be adjusted. This is beyond
the transient time period; therefore, transformer load tap changers would have
time to adjust. Transient voltage collapse due to angular instability (e.g., out-of-
step conditions) is not addressed in this report. It is assumed that transient
stability simulations for the contingencies considered have shown that the system
would be stable and meets the WECC Reliability Criteria.
2. Each member system must plan and operate its system to maintain the
minimum levels of margin specified in Table 1. Uncertainties in data,
equipment performance, and network conditions (see Section 2.3) should
be considered in the base case prior to applying Table 1.
3. Proper mixture of static and dynamic reactive power support based on the
methodology described in this report should be provided (see Section 9).
(b) The margins in Table 1 do not have to be met if (a) the local
area is radial or is a local network and (b) the contingency
under consideration does not cause voltage collapse of the
system beyond the local area.
At a minimum, each WECC member system shall conduct P-V and V-Q
analyses to ensure that the minimum required margins are met. Sole
reliance on either P-V or V-Q analysis is not sufficient to assess voltage
stability and proximity to voltage collapse. Each analysis is needed to
confirm the results of the other (i.e., P-V analysis is needed to confirm the
results of V-Q analysis and vice versa). Member systems may use either
method for general voltage stability evaluation, contingency screening, etc.;
however, voltage stability margins must be demonstrated by both P-V and
V-Q analyses. Details of the V-Q and P-V study methodologies are given
in Section 7 and Section 8, respectively.
The Councils voltage stability criteria are specified in terms of real and
reactive power margins. All member systems must provide the minimum
margins specified in Table 1 considering the uncertainties listed in Section
2.3. The margin for N-0 (base case) conditions must be greater than the
margin for Performance Level A to allow for unforeseen increases in load
or interface flows without remedial action schemes which would be
activated during contingency conditions but not during normal conditions.
Each member system must examine the items listed in Section 2.3 to
determine the required margin for its system.
Criteria noted in Table 1 apply equally to the system with all elements in
service as well as the system with one element removed and the system
readjusted. System elements include any facility, such as a generator,
transmission line, transformer, reactive power source, etc. For the
purposes of voltage stability analysis, the element, the outage of which
reduces the margin the most, should be removed from the study case and
the system readjusted following that outage. System adjustments after one
element is removed in the base case (for Performance Levels A-D
analyses) include all adjustments that can be made within 60 minutes to
bring the system to the next acceptable steady state operating condition
following the removal of the element (e.g., generation redispatch, start up
of new generation, phase shifter and tap changer adjustments, area
interchange adjustments, etc.).
The margin should be provided at all critical buses during all stressed
system conditions. Stressed cases represent worst-case conditions for
various load levels and interface flows such as
The margins in Table 1 do not have to be met if (a) the local area is
radial or is a local network and (b) the contingency under
consideration does not cause voltage collapse of the system beyond
the local area.
(1) This table applies equally to the system with all elements in service and the system with one element removed and the system
readjusted (see Section 2.2).
(2) For application of this criteria within a member system, controlled load shedding is allowed to meet Performance Level A (see
Section 2.2 for a description of provisions for application of this criteria within a member system).
(3) The list of element outages in each Performance Level is not intended to be different than the Disturbance Performance Table
in the WECC Reliability Criteria. Additional element outages have been added to this table to show more examples of
contingencies. Determination of credibility for contingencies for each Performance Level is based on the definitions used in
the existing WECC Reliability Criteria.
(4) Margin for N-0 (base case) conditions must be greater than the margin for Performance Level A.
(5) Maximum operating point on the P axis must have a MW margin equal to or greater than the values in this table as measured
from the nose point of the P-V curve for each Performance Level.
(7) Each member system should consider, as appropriate, the uncertainties in Section 2.3 to determine the required margin for its
system.
(8) The most reactive deficient bus must have adequate reactive power margin for the worst single contingency to satisfy either of
the following conditions, whichever is worse: (i) a 5% increase beyond maximum forecasted loads or (ii) a 5% increase
beyond maximum allowable interface flows. The worst single contingency is the one that causes the largest decrease in the
reactive power margin.
(*) Table 1 is an excerpt from the WSCC Reliability Criteria for Transmission System Planning in effect at the time of this
documents approval. The most current version of the Councils Table of Allowable Effects on Other Systems should be
referred to when conducting studies.
Table 1 specifies the minimum required margins for each member system.
Prior to applying Table 1, the member system should consider, as
appropriate, the uncertainties a-r [28] listed below. These uncertainties
relate to unknowns in data, equipment performance, and network
conditions. This list will be modified as member systems gain
experience in applying Table 1 and identify new uncertainties or agree
that some uncertainties are already included in Table 1.
(a) Customer real and reactive power demand greater than forecasted
(b) Approximations in studies (Planning and Operations)
(c) Outages not routinely studied on the member system
(d) Outages not routinely studied on neighboring systems
(e) Unit trips following major disturbances
(f) Lower voltage line trips following major disturbances
(g) Variations on neighboring system dispatch
(h) Large and variable reactive exchanges with neighboring systems
(I) More restrictive reactive power constraints on neighboring system
generators than planned
(j) Variations in load characteristics, especially in load power factors
(k) Risk of the next major event during a 30-minute adjustment period
(l) Not being able to readjust adequately to get back to a secure state
(m) Increases in major path flows following major contingencies due to
various factors such as on-system undervoltage load shedding
(n) On-system reactive resources not responding
(o) Excitation limiters responding prematurely
(p) Possible RAS failure
(q) Prior outages of system facilities
(r) More restrictive reactive power constraints on internal generators
than planned.
The BPA Blue Ribbon Panel report (see Appendix F) states that:
The proper amounts of static and dynamic reactive power resources are
needed to (1) supply reactive power requirements of customer demands,
(2) supply reactive power losses in transmission and distribution systems,
and (3) provide adequate system voltage support and control. The RRWG
recommends that:
1. The best method for determining the proper mixture of static and
dynamic reactive power is to conduct dynamic simulations using the
current GE or PTI programs. Member systems which already have
the capability to conduct long-term dynamic simulations should use
dynamic simulations to determine the required mixture of static and
dynamic reactive power support.
It is recommended that each member system study its loads and conduct
tests, if practical, to determine more accurate load models. The study area
should be modeled in as much detail as is feasible. It is further
recommended that bulk power delivery LTC transformers, distributions
system equivalents, voltage sensitive and dynamic loads (e.g., induction
motors) be modeled. If accurate load models are not available, loads
should be modeled as constant MVA. Accurate power factors for the loads
should also be used. Additional detailed representation of the lower
voltage distribution system is permitted. However, the system
representation used to establish voltage stability margins must be available
to all Council members for the margins to be recognized. For the purpose
of testing the margins, facility thermal ratings should be ignored to achieve
the increase in interface flows. A list of facilities which are overloaded
should be recorded, but no mitigative measures are required to resolve the
overload problems. Please refer to a document entitled, Voltage Instability
Analysis and references [12,17,20,25,26] for further information on
simulation modeling.
Each member system shall evaluate its system to identify the need and
requirements for implementing an undervoltage load shedding scheme.
The RRWG does not recommend a standard load shedding program for
the entire WECC system. Member systems should refer to the guidelines
provided in Section 10 of this report for evaluating the need for
implementation of an undervoltage load shedding program. Each member
system should investigate the applicability of undervoltage load shedding
beyond the standard contingency levels for Performance Levels A-D. The
proper level beyond the standard contingency levels (i.e., Levels A-D) to be
studied shall be determined by the member system. This load shedding
program would serve as a safety net to prevent a system blackout.
In the aftermath of the July 2 and 3, 1996 and the August 10, 1996 system-wide
disturbances occurring on the Western Interconnected electrical transmission
system, a group of WSCC members performed comprehensive assessments
culminating in two reports: the WSCC Disturbance Report for the Power System
Outages that Occurred on the Western Interconnection on July 2, 1996 and
July 3, 1996, and WSCC Disturbance Report for the Power System Outage that
Occurred on the Western Interconnection on August 10, 1996.
Recommendations from both disturbance reports identified several reliability
issues requiring further investigation.
The RRWG prepared a work plan (Work Plan) which is included as Appendix A.
The Work Plan restated the assignment, divided the work among regional areas,
identified tasks, and outlined the general methodology for accomplishing the
objectives.
The RRWG utilized information from various sources to complete the above-
stated objectives. Sources included technical papers, publications, surveys (see
Appendix B for RRWG Survey Results), BPA Blue Ribbon Panel, Operating
Capability Study Group (OCSG), feedback from experts from WSCC member
systems, and WSCC, NERC, CIGR, and IEEE guidelines/standards. The
RRWG had numerous meetings, phone conversations, and other informational
exchanges resulting in the final recommendations included in this report.
These criteria for post-transient voltage stability cover the period after the
transient oscillations have damped out to before the operator can take
manual actions and before area interchange schedules can be adjusted.
This is beyond the transient time period; therefore, transformer load tap
changers (LTCs) would have time to adjust.
If every member system adheres to these criteria, the entire Region will
have a minimum reactive power margin.
Some member systems may require more reactive power margin than the
proposed Regional Reliability Criteria.
Existing literature
The RRWG puts forward its recommendations in the spirit of the WSCC
Reliability Criteria as expressed in the fourth and fifth paragraphs of Section 2.0 -
Philosophy of the Criteria:
The RRWGs primary goal was to find a minimum criteria that would be
applicable to all member systems, i.e., to determine a minimum standard which
would apply to an individual member systems internal criteria and limit impact to
a member caused by another member. Currently, it is not required that each
individual system internal criteria conform to the Regional Reliability Criteria for
the evaluation of that systems planned performance on its own system.
However, there is an ongoing effort within the Council to change this requirement
such that the Regional Criteria would also apply to each individual internal
system. Also, due to the inherent nature of a voltage collapse phenomenon, a
voltage collapse in an area could easily spread to the neighboring interconnected
area. A shortage of reactive power in an area can load up nearby generators in a
neighboring system. Therefore, the RRWG recommends that the proposed
voltage stability criteria should apply to each individual systems internal criteria
as well as limit the impact to a member system caused by another member
system.
The criteria includes the following provisions for application to internal systems:
The margins in Table 1 do not have to be met if (a) the local area is
radial or is a local network and (b) the contingency under consideration
does not cause voltage collapse of the system beyond the local area.
Figures 5.1 and 5.2 show examples of a radial transmission system and a radial
network. Please note that the provisions apply to a bus within a radial
transmission system as shown in Figure 5.1, regardless of the number of
elements connecting that bus to the network.
Network Network
Load
Network
Load Load
69 kV 69 kV
Load
138 kV 69 kV
Load
138 kV 69 kV
230 kV
Network
The margin criteria for system contingency performance are based on results of
simulation tests. The three main factors which have been considered in defining
the criteria are as follows: (a) Performance Levels, (b) Disturbance Simulation,
and (c) The Voltage Stability Criteria Table.
The voltage stability criteria and its application will be discussed in the
following sections 5.1 and 5.2.
The RRWG concluded that the best approach for developing a margin is
based on the existing Reliability Criteria Performance Table. The
Performance Table specifies limits for voltage, frequency, and damping
based on a deterministic criteria. Therefore, it would be appropriate to
specify the margins based on a deterministic criteria and not a probabilistic
criteria. The Council is currently in the process of evaluating a probabilistic
based criteria. The RRWG members discussed setting limits for
occurrence of a voltage collapse (e.g., limit the number of voltage collapse
incidents in an area to one incident in 10 or 15 years). Since this
specification requires a probabilistic approach, it was decided to defer
setting this criteria until after the Councils Reliability Criteria has been
modified. The recommended voltage stability criteria can be modified after
the Councils Reliability Criteria has been based on a probabilistic
approach.
If imports are increased, the interface flows are also increased. The
5% increase in interface flow above the maximum allowable limit
would provide a margin for conditions where the actual flows are
higher than scheduled because of loop flow or power flow drift during
the scheduling hour due to system changes.
Member system study results show that the margins specified in Table 1
for Performance Levels A-D would be appropriate as minimums for the
region. Each member system must, however, examine the above
parameters and the list of uncertainties in Section 2.3 to determine the
required margin for its system.
Transition areas, if existent, from one to the other. (It may be fairly
straightforward to recognize the transition from one type of area to
the other. On the other hand, it may be difficult to clearly define such
a transition. It is also important to recognize that ownership
boundaries are not necessarily the transition from one area to the
other.)
5. All generators which manually control a high side remote bus must be set
at the pre-disturbance voltage at the terminal bus or local bus. Only
generators with automatic controls (i.e., no operator intervention), such as
line drop compensation, are allowed to control a high side remote bus.
Keeler SVC
Maple Valley SVC
Coulee up to 2 units
Dalles up to 6 units
John Day up to 4 units
Table 2
d) Resolve the base case with the load removed for the contingency
under consideration
In order to produce the V-Q curves, the following procedure must be followed:
2. Identify the critical bus (also referred to as the weakest bus) in the
system for this contingency. This is usually the most reactive deficient
bus. The critical bus might change with the contingency.
NOTE: If the power flow base case representing the systems post-
disturbance condition as described in Step 1 does not solve, the
6. Record the bus voltage (V) and the reactive output of the condenser (Q).
The minimum point of this curve (where dQ/dV = 0) is the critical point, i.e., all
points of the curve to the left of the minima are assumed to be unstable. The
points to the right of the minima are assumed to be stable.
If the minimum point of the V-Q curve is above the horizontal axis, the system is
reactive power deficient. Additional infeed of reactive power is required to
prevent a voltage collapse. A greater amount of reactive power is required in
order to keep a VAR margin, as quantified by the distance between the horizontal
axis and the critical point. Adequate margin must be provided to ensure system
security and reliability is maintained.
If the critical point is below the horizontal axis, the system has some margin. The
system may still be reactive deficient, depending on the desired margin, and
maintaining acceptable post-transient voltages. Additional infeed of reactive
power is required if a greater margin is desired.
1. Start with base cases that simulate reasonably adverse conditions for
different seasons and load levels. (The load levels studied may, but not
necessarily, be at the critical point on the P-V curve.)
Another method to determine the weakest bus is by monitoring the dV/dQ from
the Jacobian matrix at buses throughout the system when the P-V curve is
calculated (see Section 8 for further information on P-V methodology). The bus
that has the largest rate of change of the dV/dQ before collapse (i.e., the nose)
is the weakest bus.
Sections 7.2 and 7.3 describe two study methodologies: one for load area and
one for interfaces.
The following methodology should be followed for studies involving load areas:
1. Start with an appropriate base case to represent worst case conditions for
the load area of interest. Scale the loads to a 1 in 2 year occurrence load
forecast and adjust resources appropriately. Load forecasts with a lower
probability (e.g., 1 in 10 year occurrence, 1 in 20 year, etc.) could also be
used.
2. Assume constant MVA loads unless more accurate load models are
known. If load models other than constant MVA are used, the load voltage
regulators/transformers in the distribution system must be modeled and
active in the base case. The distribution system representation must be
available to all WECC members for the resulting margin to be recognized.
Accurate power factors for the loads should also be used.
6. Conduct V-Q studies for each of the worst contingencies identified for
Performance Levels A-D to determine if the reactive power margin meets
the required margin. The required margin includes the minimum margins
shown in Table 1 considering the list of uncertainties in Section 2.3.
7. If the required margin is not met, determine if there are additional existing
or planned facilities that are designed and expected to operate successfully
during the post-transient time frame. Re-run the case adding these
facilities to the post-transient procedure listed above to determine if the
required margin is met. If the required margin is still not met, additional
facilities or implementation of appropriate remedial action schemes would
be required.
1. Start with an appropriate base case to represent worst case conditions for
the interface of interest. The interface must be at its maximum rating and
the worst load condition that the interface is rated for must be modeled.
2. Assume constant MVA loads unless more accurate load models are
known. If load models other than constant MVA are used, the load voltage
regulators/transformers in the distribution system must be modeled and
active in the base case. The distribution system representation must be
available to all WECC members for the resulting margin to be recognized.
Accurate power factors for the loads should also be used.
4. Automatic adjustments, which would occur within 0-3 minutes, are allowed
for increasing the interface flow. These adjustments include tap changer
and phase shifter adjustments, switching of facilities, etc. Automatic
adjustments within 3 minutes are allowed since interface flows can change
very quickly due to changes in operating conditions and, generally
speaking, manual system adjustments are not made as frequently as for
the case with load increases.
7. Conduct V-Q studies to determine if the reactive power margin meets the
required margin. The required margin includes the minimum margins
shown in Table 1 considering the list of uncertainties described in
Section 2.3.
8. If the required margin is not met, determine if there are additional existing
or planned facilities that are designed and expected to operate successfully
during the post-transient time frame. Re-run the case adding these
facilities to the post-transient procedure listed above to determine if the
required margin is met. If the required margin is still not met, additional
facilities or implementation of appropriate remedial action schemes would
be required.
For V-Q studies, minimum reactive power margin will be specific to the
system and conditions under study. The minimum reactive power margin
for Performance Level A shall be determined for each system by finding
the maximum change in reactive power margin at the most critical bus in
the study area as follows:
(a) Determine the worst contingency for Level A for the area under
study. This can be done by conducting numerous post-transient
contingency studies involving large disturbances during worst-case
conditions such as peak load and/or maximum import conditions.
(c) Increase the area load or the flow on the critical interface by 5% and
develop the corresponding V-Q curve for Performance Level A
contingency.
The change in the margin is the amount of margin that must be provided for
Performance Level A. An example is given in the following paragraphs.
The first task is to determine how much margin is needed for V-Q analysis for the
worst single contingency. Assume a system is in steady state worst-case
conditions (e.g., peak loads and/or maximum interface flow levels) with all
elements in-service at time t=0. Further assume that the worst single
contingency and the most critical bus for this system have been already identified
by conducting numerous contingency and V-Q analyses. The margin must cover
the worst contingency for Performance Level A and a 5% increase in load or
interface levels.
Consider the V-Q curve for the N-0 conditions as shown in Figure 7.2. Assume
that the V-Q curve for the worst N-1 results in a reduction of margin by 500
MVAR. Furthermore, assume that the base load (i.e., both real and reactive
power load) is increased by 5% and the V-Q curve is produced for the worst N-1
case. Suppose that this results in a further reduction of 300 MVAR in reactive
power reserve as measured at the nose point and compared with the V-Q curve
for the worst N-1 case with base load.
Worst N - 1
Contingency
Q
Base Load or
Interface Flow
N-0
Base Load or
Interface Flow
300
0.95 1.0 V
500
Figure 7.2 - V-Q Curve Test for Determination of Reactive Power Margin
Therefore, the total required margin is 300 MVAR for the worst single contingency
without system adjustments. This implies that after a system experiences an N-1
contingency, at least 300 MVAR of margin must be available. The system is
clearly deficient in reactive power and has a potential to collapse. After installing
reactive power support, the V-Q curves must be reproduced to provide the
necessary positive margin as shown in Figure 7.3. The margin covers the worst
single contingency and a 5% load forecast or interface flow uncertainty. If
capacitors are added to provide the required margin, an adequate amount must
be added to take into account the relationship between the capacitor output and
square of the voltage applied to the capacitors. For example, if the required
margin of 300 MVAR is at a voltage collapse point of 0.9 pu, about 370
(300/0.92 =370) MVAR of capacitors would be needed to provide the margin.
0.95 1.0
300
V
Worst N - 1
Contingency
Base Load or
Interface Flow + 5%
Worst N - 1
Contingency
Base Load or
Q Interface Flow
N-0
Base Load or
Interface Flow
400
0.95 1.05
V
600
Test
Bus
Generator
~
100
50 MVA
MVARS
Begin a P-V analysis with an approved WECC power flow base case (i.e., with at
least a 1 in 2 year occurrence load forecast) to ensure that the region external to
the study area is represented in a reasonably accurate manner. An area
susceptible to voltage collapse can be identified by a power flow contingency
analysis. Cases that fail to converge to a solution or that exhibit large post-
transient voltage deviations are typically in or near a voltage unstable operating
point, respectively. If the power flow program has the feature that can monitor
dV/dQ from the Jacobian matrix during the P-V run, these quantities can provide
information on the buses where the collapse will begin. The bus that has the
largest rate of change of the dV/dQ before collapse (i.e., the nose) is the
weakest bus (see Section 7 for further information about locating the weakest
bus).
Of the three types of load representation, constant MVA (a.k.a. constant power),
constant current and constant impedance, constant MVA typically results in the
most pessimistic point-of-collapse in a P-V analysis. A constant MVA load
representation approximates the action of distribution system voltage regulating
devices and, therefore, should be used unless more accurate load
representations are known. If more accurate load representations are used, they
should be modeled on the low side of load-serving transformers and the detailed
modeling of voltage regulators on those transformers added to the system
representation. Additional detailed representation of the lower voltage distribution
system is also permitted. However, the system representation used to establish
A full P-V curve (see Figure 8.1) can be produced by two methods. The first is by
increasing the loads in the study region and increasing external generation. The
second is by increasing flows across an interface (i.e., shifting generation from
the receiving region to the external region). External areas that are resource
constrained are permitted the use of a fictitious generator, solely for the purpose
of establishing power margin. Good engineering judgment should be used in
placing such a generator, and the generator should not supply reactive power.
The following procedures cover these two methods of stressing a region to
generate a complete P-V curve from low loads or low interface flows to high loads
or high interface flows, respectively. Separate procedures are also provided and
are to be used when the primary analysis method is V-Q and P-V checks are
needed to ensure that the power margin is met.
The procedure for full P-V curve development, P-V tests, and determination of
real power margin are discussed in the following sections 8.1, 8.2, 8.3, 8.4, and
8.5.
2. Model the loads in the study area initially at a level of approximately 20% of
the expected peak load. This will provide the full benefit of P-V analysis, with
the development of the P-V curve at levels below the operating points for
which problems could be anticipated. Generation external to the study area
should be reduced to match the scaled down load levels in the study area.
As loads are scaled up in the study area, the effects of increased load
requirements on the study region's voltage profile will be captured.
Voltage
Pre-contingency
Post-contingency case
case
Real Power
3. Set the internal study area generation to a constant level of the on-line units.
The real power output of the internal generators should remain unchanged
during the P-V analysis. The reactive power capability of each of the
generating units should represent the unit's capability, and the reactive
power output of each unit should be allowed to adjust as the P-V analysis
progresses. Voltage collapse will occur in the study region after the VAR
capability in the study region is depleted.
4. Choose the bus or buses in the study area at which the voltages will be
monitored as the power transfers into the study area are increased. As an
initial investigation of a region for voltage instability, the engineer should
select several buses to monitor. The monitored voltages are the y-axis data
of a P-V curve. See sections 7 and 8 regarding methods of identifying buses
to monitor.
5. Determine (a) if the x-axis data will be load or interface flows, and (b) if the
units will be MW or MVA (see Figure 8.1). If an interface path is used, it
should be defined in a manner that measures all imports to the receiving
region. A partial interface definition that allows imports into the receiving
region over an unmonitored branch is incomplete. Choosing the x-axis to be
study area load in MW is a good starting point.
7. Solve the initial power flow case representing a low receiving area load for
the performance level being studied using the post-transient methodology
described in Section 6.
8. Record the bus voltages at the monitored buses, and the load level or
interface transfer level at which the power flow case was solved.
9. Scale loads and external generation up to match the load increase. The
load increases can be larger at lower load levels than at higher load levels,
which are near the point of collapse. Ensure that loads are scaled up in the
neighboring systems if they have similar climatic or geographic
characteristics to the system under study. Initially, a load increase equal to
the starting load level in the study region should be effective. If the power
flow case fails to converge to a solution after a load increase, return to the
last solved case, and scale up the loads by one-half or one-fourth of the
previous attempt.
NOTE: When the load level reaches the starting load level, for the next 5%
load increase, only automatic and manual adjustments, which
would occur within 30 minutes, are allowed for increasing the load.
These adjustments include generation dispatch, tap changer and
phase shifter adjustments, switching of devices, etc.
10. The results of the P-V analysis could indicate that the voltage profile of a
region is significantly lower than acceptable operating conditions at the point-
of-collapse. In such cases, the limit of the system could be determined by
other voltage criteria, such as post-transient voltage deviation or the lower
limit of acceptable operating voltage. However, in some receiving regions,
typically regions with a high degree of shunt compensation, the point-of-
collapse will occur at or near bus voltages that appear acceptable. For these
cases, the system should be designed with some operating margin from the
point of collapse.
The engineer starts with the WECC base case previously used in the power flow
analysis, and scales loads in the study area to 20 MW at a constant power factor
(It is assumed that the neighboring systems do not have a climatic or geographic
characteristics similar to the study area.). Generation external to the study area
is reduced by 80 MW. The internal generation is set at 25 MW with the
appropriate VAR capability. The contingency is simulated and the power flow
case is solved. At the 20 MW load level, the voltages at the monitored buses are
recorded.
The loads in the study area and external generation are scaled up by 20 MW and
the case is solved. At the 40 MW load level, the voltages at the monitored buses
are recorded. The loads in the study area and external generation are then
scaled up by another 20 MW, the case is solved, and the monitored bus voltages
are recorded at the 60 MW load level. This process is also repeated for the 80
MW load level. However, at the 100 MW load level, the case fails to converge.
At this point, the engineer returns to the 80 MW case, and this time scales study
area loads and external generation up by only 10 MW, the case is solved, and the
monitored bus voltages are recorded at the 90 MW load level. This process is
repeated for the 100 MW load level. However, at the 110 MW load level, the
case fails to converge again.
The engineer returns to the 100 MW case, and this time scales study area loads
and external generation up by only 5 MW, solves the case, and records the bus
voltages. The process is repeated at the 110 MW load level, but this case also
fails to converge. The engineer returns to the 105 MW case and scales up study
area loads and external generation by 2.5 MW. This case is solved and the
monitored bus voltage are recorded at the 107.5 MW load level. The loads and
external generation are then scaled up another 2.5 MW and the case fails to
converge. The engineer returns to the 107.5 MW load level and scales study
area loads and external generation up by 1.25 MW, but this case also fails to
converge. At this point the engineer decides that the point-of-collapse is
determined to a reasonable accuracy, at 107.5 MW.
Based on this analysis, the engineer knows that Level A is met for the simulated
contingency at historical peak loads, because the power margin is 7.5%.
2. The loads in the study area should be at highest possible levels and realistic
power factors. The loads represented should be consistent with the
projections for the season and year being evaluated for conformance with
the performance levels in Table 1. As generation is shifted from the study
area to the external area, the effects of increased power transfers across a
monitored path on the study region's voltage profile will be captured.
3. Set the internal study area generation to a higher than normal level. The
reactive power capability of each of the generating units should represent the
unit's capability, and the reactive power output of each unit should be
allowed to adjust as the P-V analysis progresses. Voltage collapse will occur
in the study region after the VAR capability is depleted.
4. Choose the bus or buses in the study area at which the voltages will be
monitored as the power transfers into the study area are increased. As an
initial investigation of a region for voltage instability, the engineer should
select several buses to monitor. The monitored voltages are the y-axis data
of a P-V curve. See Sections 7 and 8 regarding methods of identifying
buses to monitor.
5. The x-axis data will be interface flows, measured in either MW or MVA (see
Figure 8.1). The interface path(s) should be defined in a manner that
measures all imports to the receiving region. A partial interface definition
that allows imports into the receiving region over an unmonitored branch is
incomplete.
7. Solve the initial power flow case representing a low path flow (e.g., 20% of
estimated maximum path rating) for the performance level being studied.
8. Record (a) the bus voltages at the monitored buses, and (b) the interface
transfer level at which the power flow case was solved.
9. Shift generation from the study area to the external area. The generation
shifts can be larger at lower path flows than at higher path flows, which are
near the point of collapse. Initially, a generation shift of 10% of the study
region generation should be effective. If the power flow case fails to
converge to a solution after a generation shift, return to the last solved case,
and reduce the shift by one-half or one-fourth of the previous attempt.
NOTE: When the interface flow reaches the starting interface flow level in
the base case, for the next 5% flow increase, only automatic
adjustments which would occur within 3 minutes, are allowed for
increasing the interface. These adjustments include generation
dispatch, tap changer and phase shifter adjustments, switching of
devices, etc.
10. The results of the P-V analysis will often indicate that the voltage profile of a
region will be significantly lower than acceptable operating conditions at the
point-of-collapse. In such cases, the limit of the system could be determined
by other voltage criteria, such as post-transient voltage deviation or the lower
limit of acceptable operating voltage. However, in some receiving regions,
typically regions with a high degree of shunt compensation, the point-of-
collapse will occur at or near bus voltages that appear acceptable. For these
cases, the system should be designed with some operating margin from the
point of collapse.
The generation in the study area and external generation are shifted by 100 MW
and the case is solved. At the 300 MW load level, the voltages at the monitored
buses are recorded. This process is repeated for the 400, 500, 600, 700, and
800 MW transfer path flow level. However, at the 900 MW load level, the case
fails to converge. At this point, the engineer returns to the 800 MW flow case,
and this time shifts only 50 MW of generation, the case is solved, and the
monitored bus voltages are recorded at the 850 MW transfer path flow level. This
process is repeated for the 900 MW load level; however, the case fails to
converge again.
The engineer returns to the 850 MW case, and this time 25 MW of generation is
shifted, the case is solved, and the bus voltages recorded. The process is
repeated at the 900 MW load level, but this case also fails to converge.
Therefore, the engineer returns to the 875 MW case and scales up study area
loads and external generation by 12.5 MW. This case is solved and the
monitored bus voltage is recorded at the 887.5 MW flow level. Another 12.5 MW
generation shift is made and the case fails to converge. The engineer returns to
the 887.5 MW flow level and shifts 6.25 MW, but this case also fails to converge.
At this point the engineer decides that the point-of-collapse is determined to a
reasonable accuracy, at 887.5 MW.
Based on this analysis, the engineer knows that Level A is more constrained by
voltage collapse than by the previous power flow and dynamic analyses.
The WECC accepted methodology for conducting a P-V test for studies involving
load areas is described below. The purpose of this test is not to develop a full
P-V curve as described in Sections 8.1 and 8.2 but to ensure that the margins in
Table 1 are met. The following methodology must be followed to test the real
power margin.
1. Start with an appropriate base case to represent worst case conditions for
the load area of interest. Scale the loads to a 1 in 2 year occurrence load
forecast and adjust generation and imports accordingly. Load forecasts with
2. Identify the critical bus (also referred to as the weakest bus) in the system
for this contingency (see Sections 7 and 8 for further information about
locating the weakest bus).
3. Increase loads within a well defined load region in small steps (usually 1% of
load per step). Ensure that loads are also scaled up in the neighboring
systems if they have similar climatic or geographic characteristics. Increase
generation from remote area(s) outside the study area which supply power
to the load region to stress interface paths in the load serving area.
4. Assume constant MVA loads unless more accurate load models are known.
If load models other than constant MVA are used, the load voltage
regulators/transformers in the distribution system must be modeled and
active in the base case. The distribution system representation must be
available to all WECC members for the resulting margin to be recognized.
Accurate power factors for the loads should also be used.
6. Apply the contingency for the appropriate Performance Level using the post-
transient methodology described in Section 6 and solve the power flow case.
8. Repeat steps 3 to 7 until the nose point has been reached or the case does
not solve.
10. The maximum operating points must have a margin equal to greater than
the required margin for the performance level under consideration as
measured from the nose point on the P-V curve. For example, the
maximum operating point for Performance Level A is defined as the load
level at the nose point multiplied by 0.95 (This represents a 5% margin for
11. If the required margin is not met, determine if there are additional existing or
planned facilities that are designed and expected to operate successfully
during the post-transient time frame. Re-run the case adding these facilities
to determine if the required margin is met. If the required margin is still not
met, additional facilities or implementation of appropriate remedial action
schemes would be required.
12. Conduct P-V analysis for each of the worst outages identified for
Performance Levels A through D (following steps 2-11 above).
The WECC accepted methodology for completing a P-V analysis for studies
involving interfaces for a well-bounded area is described below. The purpose of
this test is not to develop a full P-V curve as described in Sections 8.1 and 8.2 but
to ensure that the margins in Table 1 are met. The following methodology must
be followed to test the margin.
1. Start with an appropriate base case to represent worst case conditions for
the interface of interest. The interface must be at its maximum rating and
the worst load condition that the interface is rated for must be modeled.
3. Assume constant MVA loads unless more accurate load models are known.
If load models other than constant MVA are used, the load voltage
regulators/transformers in the distribution system must be modeled and
active in the base case. The distribution system representation must be
available to all WECC members for the resulting margin to be recognized.
Accurate power factors for the loads should also be used.
5. Automatic adjustments which would occur within 3 minutes are allowed for
increasing the interface. These adjustments include tap changer and phase
shifter adjustments, switching of devices, etc. Individual curves should be
generated for each automatic adjustment. Individual curves are not needed
6. Apply the contingency for the appropriate Performance Level using the post-
transient methodology described in Section 6 and solve the power flow case.
8. Repeat steps 3 to 7 until the nose point has been reached or the case does
not solve.
10. The maximum operating points must have a margin equal to greater than
the required margin for the performance level under consideration as
measured from the nose point on the P-V Curve. For example, the
maximum operating point for Performance Level A is defined as the
interface flow at the nose point multiplied by 0.95 (This represents a 5%
margin for Level A). The required margin includes the minimum margins
shown in Table 1 considering the list of uncertainties described in Section
2.3.
11. If the required margin is not met, determine if there are additional existing or
planned facilities that are designed and expected to operate successfully
during the post-transient time frame. Re-run the case adding these facilities
to determine if the required margin is met. If the required margin is still not
met, additional facilities or implementation of appropriate remedial action
schemes would be required.
12. Conduct P-V analysis for each of the worst outages identified for
Performance Levels A through D (following steps 2-11 above).
Similar to the V-Q methodology, tests should be done to determine the required
amount of MW margin on a P-V curve. For example, consider the P-V curve for
the N-0 conditions as shown in Figure 8.2. Assume that the P-V curve for the
worst N-1 case results in a reduction of margin by 500 MW as measured from
nose point of the P-V curve for the N-0 case to the nose point of the P-V curve for
the N-1 case (distance between point 1 and point 2 as shown in Figure 8.2).
Furthermore, assume that a 5% margin on the P-V curve for the worst N-1 case
is considered to allow for load forecast uncertainty. The 5% is based on the load
level measured at point 2. This results in limiting the load level or interface flow
to a value shown as point 3 in Figure 8.2. Distance between point 2 and point 3
represents the 5% margin for the worst N-1 contingency.
Suppose that point 1 and point 2 represent a maximum load level or interface
flow of 2000 MW and 1500 MW, respectively. This results in limiting the load
level or interface flows to 1425 MW (0.95 * 1500 = 1425 MW).
N -1
Contingency
3
75 2
MW 500
MW 1
PerformanceLevel A
V PerformanceLevel B
PerformanceLevel C N- 0
PerformanceLevel D
>0 2.5
2.5%
%
5%
8 76 54 3 2 1
P
Figure 8.3 - P-V Curve Test for Determination of Real Power Margin
If manual or automatic adjustments are normally made to increase the flows into
a load area (within 30 minutes), or automatic adjustments are normally made to
increase interface flows (within 3 minutes), additional curves should be generated
for each system adjustment, as appropriate.
The preceding sections of this report propose voltage stability criteria and
methodologies to ensure that an adequate amount of reactive power would be
available. If studies show that additional sources of reactive power would be
needed to meet the criteria, a further study may be needed to determine the
proper mixture of static (e.g., capacitors) and dynamic (e.g., SVCs, synchronous
condensers, or generators) sources of reactive power.
Dynamic reactive power support and voltage control are essential during
power system disturbances. Synchronous generators, synchronous
condensers, and static var compensators (SVCs and STATCOMs) can
provide dynamic support. Transmission line charging and series and shunt
capacitors are also sources of reactive support, but are static sources.
Standards
9.1 Discussion
In order to further illustrate the need for determination of the proper mixture, an
example is given. Suppose that voltage stability studies show that a system
needs 400 MVAR of additional reactive power support. The reactive power
deficiency can be solved by several options such as:
2. Install generation
3. Install SVCs
Load shedding is also an option for Performance Level A contingencies but only
for internal systems. For the purpose of this example, it is assumed that load
shedding is not considered. Therefore, option 6 would likely be the least
expensive option. If any option(s) in 1-5 is selected, no further studies are
needed for determination of the proper mixture. However, if it is decided to install
automatically switched shunt capacitors, a further study may be needed to
determine how much of the required 400 MVAR can be supplied by automatically
switched shunt capacitors. The potential problems with installing 400 MVAR of
capacitors are as follows:
If the proposed voltage stability criteria are met with no switched reactive power
sources after the contingency during the most stressed conditions, no further
studies are needed. Normally, this would entail demonstrating that margins are
met with constant MVA loads. If the proposed criteria are met using switched
reactive power sources and load models which provide relief for depressed
voltages, further studies are needed to establish that the system will meet the
criteria as voltages are restored to the load.
Worst N - 1 Contingency,
Base Load or
Interface Flow + 5%
N - 0, Base Load
or Interface Flow
Static 200
100
Dynamic
100
500
Figure 9.1- V-Q Curve Test for Determination of Mixture of Static and
Dynamic Reactive Power Sources
Figure 9.1 shows that the system requires a margin of 300 MVAR. In order to
provide the required margin, a total of 400 MVAR of reactive power must be
added. To determine the mixture of static and dynamic reactive support needed,
the Q-V curve analysis for the worst N-1 contingency and additional 5% of load is
repeated by using a post-contingency, short-term load model rather than a long-
term load model. The short-term load model may represent a lower load level
due to reduced distribution voltages immediately after a contingency before slow
control actions (e.g., transformer tap changers) have taken place. The post-
transient procedures described in Section 6 for simulating the disturbance for
each performance level should be followed except that all slow control actions
(e.g., switching of capacitors, tap changer adjustments, etc.) should be turned off
in the simulation. For example, the voltage at a particular distribution bus may be
1.0 pu before the contingency, 0.90 pu immediately after a contingency, and 0.95
pu after control actions have taken place. The load magnitude at this bus could
vary as the voltage changes. The long-term load model would represent the full
load after all control actions such as tap changers have taken place.
As described in this report each member system must develop load models for its
system. A load model may look something like this:
P = P0 (V/V0)a Q= Q0 (V/V0)b
P = a P0 V2 + b P0 V + c P0
Q = d Q0 V2 + e Q0 V + f Q0
Where a, b, c, d, e, and f are per unit multipliers for various types of load
(i.e., constant power, constant current, and constant impedance).
After the load models have been developed, they can be used to estimate the
values of P and Q by using the voltage magnitude immediately after a
contingency. The short-term load model can be used in the power flow programs
such as GE or PTI for the methodology described above to develop the V-Q
curve immediately after a contingency. Automatic transformer tap changers must
be locked at pre-contingency settings. This is important since tap changers
would attempt to restore voltages to pre-contingency values which is not desired
for this analysis.
The dashed V-Q curve shown in Figure 9.1 uses a short-term load model.
Figure 9.1 shows that at least 200 MVAR of the reactive power support must be
provided in dynamic form such as SVCs, synchronous condensers, etc. The
remaining 200 MVAR can be provided by mechanically switched capacitors with
automatic controls.
9.2 Recommendations
1. The best method for determining the proper mixture of static and
dynamic reactive power is to conduct dynamic simulations using the
current GE or PTI programs. Member systems which already have the
capability to conduct long-term dynamic simulations should use dynamic
simulations to determine the required mixture of static and dynamic
reactive power support.
Throughout this section and when studying voltage limited systems, undervoltage
load shedding should be ranked among many options available to recover
reactive power margin. Other options include shunt and series capacitor
additions, automatic reactor and capacitor switching, MSCs, maximizing local
generator reactive capacity, generator line drop compensation, and blocking or
coordinating load-tap changing.
Since voltage collapse can occur suddenly, there may not be sufficient time for
operator action to stabilize the system. Therefore, a load shedding scheme that
is automatically activated as a result of undervoltage system conditions may be
an effective means of stabilizing the system and mitigate the effects of a voltage
collapse. Undervoltage load shedding schemes should be coordinated with other
system measures used to disconnect load, such as remedial action direct load
tripping.
Automatic undervoltage load shedding can be used with large areas that have
strong transmission systems as a guard against voltage collapse in the event that
multiple contingencies and extreme conditions should occur. When applied to
low probability, multiple contingency events, undervoltage load shedding provides
a low-cost means of preventing widespread system collapse.
An example of small area undervoltage load shedding is where a new 230 kV line
and 230-115 kV transformer was delayed because of permitting and right-of-way
issues. Undervoltage relays were employed, with a transfer signal that indicated
loss of the transmission element. If the transmission element was lost from
service, and undervoltage was detected, loads at the far end of the system were
automatically tripped, until a local combustion turbine could be started. This was
used for two years, until the reinforcement could be placed in-service.
the amount of required load shedding varies from one area to another
and must be determined locally by conducting studies; and
too much load shedding could result in generator instability.
10.1 Guidelines
10.2 Methodologies
Load models are recommended in papers by Carson Taylor, reference 25, and
IEEE, reference 26. The models represent the voltage and time sensitivity of
aggregate loads as represented at the distribution substation. Appendix G
contains NERC Planning Standards on Undervoltage Load Shedding.
The capability to use P-V, V-Q, or other methods in real-time operation does not
currently exist. In the absence of any on-line software program, the only easy
way for operators to measure margin, in real-time, at this time, is by measuring
the available reactive power reserve from various sources. Reactive power
reserves of generators, synchronous condensers, and SVCs are sensitive
indicators of voltage stability. If the available reactive power reserve from a
generator is diminished such that its field current limit is reached, the field current
of nearby generators could also reach their limits resulting in voltage collapse.
Requirements
Guides
Criteria
After taking all other remedial steps, a system or control area whose integrity is
in jeopardy due to insufficient generation or transmission capacity shall shed
customer load rather than risk an uncontrolled failure of components of the
interconnection (C.III.C.)
Requirements
1. Plans for automatic load shedding. Each system shall establish plans
for automatic load shedding (V.D.R.1.)
1.2.1. Load shedding steps. Automatic load shedding shall be in steps related
to one or more of the following: frequency, rate of frequency decay,
voltage level, rate of voltage decay or power flow levels. (V.D.R.1.2.1)
1.2.2. Minimizing risk. The load shed in each step shall be established to
minimize the risk of further uncontrolled separation, loss of generation, or
system shutdown. (V.D.R.1.2.2.)
10. HVDC. Entities with HVDC transmission facilities should use the
reactive capabilities of converter terminal equipment for voltage
control.
Rky
CA NW Mtn DSW
Abbas Brian Frank Robert
If the answer to the above question is yes, can this margin be applied to all
member systems? If the answer is no then how should the margin be
specified?
(h) "Planning Against Voltage Collapse", Electra, No. 111, pp. 55-
75. (Frank, Kevin)
2. Conduct a survey
-Find out what the member systems are using in terms of voltage
collapse criteria (margin, contingency, methodology, modeling, etc.)
7. Review work done by "Reactive Study Review Group" of the Northwest and
prepare a summary. (Brian, Andy, Joe to be completed by December 20,
1996)
10. Determine requirements for setting a reactive reserve margin and prepare a
summary report (All, to be completed by February 28, 1997):
-Amount
-How to measure?
11. Investigate what methodology to use for measuring margin and prepare a
summary report (All, to be completed by February 28, 1997):
-Feasibility
12. Determine what tools to use (All, prepare a summary report by February 28,
1997):
-OPF
-VSTAB
-Dynamic (ETMSP)
13. Investigate what modeling should be used for voltage collapse analysis and
prepare a summary report (All, to be completed by February 28, 1997)
14. Find out what other member systems outside the Council are doing with
respect to our objectives.
15. Compare requirements for setting voltage collapse criteria vs. transient
stability criteria and prepare a summary report. (All, to be completed by
February 28, 1997)
17. Provide answers to the following questions (All, prepare a summary report by
March 28, 1996):
If the answer to the above question is yes, then can this margin be applied to
all member systems? If the answer is no then how should the margin be
specified?
18. Prepare a draft document summarizing the results of our investigation and
recommendation and send for review. (All to be completed by April 15, 1996)
(Frank, Kevin)
(Frank, Kevin)
12. "Voltage Stability Analysis Using Generic Dynamic Load Models", W. Xu, Y.
Mansour, B.C. Hydro
Planning:
Static:
Quasi-dynamic:
Dynamic:
Operating:
Others
Deterministic specification
Probabilistic specification
Definition of credibility
The following items were considered for determining the amount of margin:
Fixed
Variable
% of load
% of critical flow
Single contingency
Double contingency
First contingency:
VAR Support
Nomograms
Load Shedding
Load shedding
Feasibility
Settings to use
Equity
SURVEY RESULTS
RRWG WSCC
Survey Summary
V-Q analysis 6 13
Other 3 1
None 2 -
P-V analysis 1 2
Nomograms - 1
RRWG WSCC
Survey Summary
500 MVAR 1 3
400 MVAR 1 1
250 MVAR - 1
200 MVAR 1 -
35 MVAR 1 1
50 MVAR 2 1
None 5 4
4% 1 1
5-10% - 1
5 MW 1 -
Load Shed 1 -
Other - 1
N-1 3 4
N-2 4 4
Credible Multiple 2 -
None 5 9
Voltage Criteria
None 7 8
WSCC post-transient 3 2
15 kV margin, 480 kV 1 -
2% dip 2 2
5% margin 1 2
10% dip - 4
Yes : 4
Qualified Yes: 7
No Opinion: 2
No 1
An N-2 criteria should be developed for voltage instability. N-2 criteria may
be appropriate because N-2 outages are more severe and they do occur.
Therefore, planning for an N-2 outage will ensure a safe and reliable
operating region. The N-2 criteria if included must have a less stringent
requirement than the N-1 criteria. The important thing to remember is that
the N-2 outage must be a credible disturbance and must be the initiating
event.
A criterion of N-2 should be used. The criteria for a N-1 and N-2 may differ.
Yes, the WSCC WATT criteria is N-2, the VAR criteria should be similar.
N-2 is not a criteria for voltage instability in our area. Voltage instability can
occur in our area for an N-1 condition. Voltage instability may also occur
under N-0 conditions, for some systems, as load ramps and insufficient
amount of reactive compensation is dispatched to follow the increased
load.
If the suggestion is to allow for a less stringent criteria for a less probable
outage, this only seems appropriate as a corollary to the damping criteria.
With voltage collapse eminent under single or multiple contingency, the
amount of margin should be greater than zero and possibly equal to zero
for the least probable contingency as is the damping criteria. The
credibility of the contingency should be what governs the amount of
margin.
Partially agree. We agree for the common-mode N-2 outages, not for the
independent events. Since the N-2 outages are area and disturbance
dependent, the local criterion should be met. It is hard to have a generic
cross-region N-2 criteria in WSCC system.
We dont believe that N-2 is that bad of a criteria if you consider loss of a
VAR source as one of the elements, i.e., loss of a VAR source and a line.
No Opinion: 5
Qualified Yes: 5
Yes : 2
No 2
Maybe in the sense that the probability of a three phase fault is low and
that it is a more severe disturbance compared to a single phase
disturbance.
It has been our experience that an N-2 event is a more probable cause
of voltage instability.
I suppose there are similarities, but the fact is that we have to define
some principles to which we will design our system. Certainly, the
probability of a three-phase fault is remote, but the issue is that we
agreed to design our systems to that level. Even though the probability
is low, I believe that the transient stability criterion is acceptable.
Yes, but the mixing of criteria with the probability of an outage does not
make sense here. This answer addressed only the probabilities. The
probability of a multiple contingency could be similar to that of a single
contingency. Once the multiple contingency has occurred, the
probability of another occurrence would increase.
Not studied 3
Load shed 5
No action needed 6
Yes 2
Qualified yes: 5
No 6
No opinion 1
Yes 7
No 5
No answer 2
8. How would you characterize the voltage instability problems you might
have identified? (e.g., fast/slow decay?)
Fast 2
Slow 5
N/A 7
None 2
V-Q 1
I think that we need to proceed wisely into this endeavor. The trend is to
define some VAR margin, when the important factor is to simply know that
the amount of VARS on the system are adequate. The P-V method of
analysis has much more meaning to a system operator.
There is a need to manage Voltage Security of the system based on the July 2
and August 10, 1996, system disturbances. Voltage Security assessment
involves monitoring of the reserve available in dynamic system sources like
generators and SVCs. The Reactive Power (MVAR) Monitor was developed to
enable Dispatchers to continuously assess the system and to alarm when
pre-determined limits are reached. A voltage security index is computed which
weights projects according to their sensitivity to change and their size. The index
is calculated so that it will be zero [0.0] if all plants are at or above their
continuous reactive power limit. [Sort of like the gas gauge on a car]. The
weighting is so a large plant close to the voltage control area will have a high
weight factor, and therefore affect the index more than remote plants would.
Initially the sensitivities will be set using off-line power flow studies, but long term
plans are to have the sensitivities calculated near real-time by on-line power flow
studies.
PROCEDURE
The Reactive Power (MVAR) Monitor is one of three monitors described in DSO
303 and is to be used as a means of monitoring dynamic reactive reserve and
triggering manual system changes to maintain dynamic reserve. If sufficient
reserve cannot be maintained, DSO 303 gives guidance and ultimate authority to
cut schedules if necessary. The calculated index is a more sophisticated
measure of reactive reserve than a flat MVAR amount. The calculated index
takes location and system conditions into consideration.
ALARMS
Alarming is in two modes, color change and an audible alarm. The limit [an
enterable field] is set manually to give warning before running reserve to the
bottom. When we violate the calculated index as displayed on page 1, the
graphic will change from GREEN to RED. The second graphics showing the
calculated index is displayed in yellow. When the limit is violated, an AUDIBLE
alarm will sound on the SCADA-R terminal. When the alarm sounds, click on the
ALL ALARM icon in the APPLIST icon box at the left of the screen. Click on the
SILENCE button at the upper right to silence the alarm. You will see a flashing
message on the alarm page displaying REACTIVE [MVAR] MONITOR
VIOLATION DETECTED. The alarm will again be initiated when the actual
calculated index goes above the limit to allow for going back to pre-violation
limits.
DISCUSSION
The index is computed for each critical voltage control area. It is planned to have
indices for the following areas:
The PSST will be used for the platform initially, but when SCADA-R is
implemented at Dittmer, the application will migrate to that platform.
NOTE: This monitor does not predict transient voltage stability, only medium and
long-term dynamic voltage stability.
BC Hydro [3], Florida Power and Light [4], and Swedish utilities [5,6] use
generator reactive power level, along with HV or EHV voltage magnitude,
for automatic load shedding and other countermeasures.
We are proposing two main displays. The first is bar charts showing power
plant reactive power output and capability. Operators can see which plants
are approaching reactive power limits and take appropriate actions such as
capacitor bank switching to reduce output. If plants are at or above their
continuous reactive power capability, emergency actions may be
necessary.
The second main display is bar chants showing VSA indices for different
areas (Puget Sound area, Portland area, overall west side, and east side).
The indices (scale of 0 to 100) are computed for each area based on
weighted values of generator and SVC reactive power output divided by
weighted generator and SVC reactive power capability. The basic idea is to
avoid current limiting at critical plants. Higher weights are given to nearby
plants and to large plants, meaning that current limiting at a large nearby
plant will significantly reduce voltage security. Weights are computed from
power flow program sensitivity analysis. Eventually, weights can be
computed on-line using the on-line power flow program.
Other displays provide more details such as individual power plant line real
and reactive power and trends of indices. The plants monitored are:
References
1. P. Nedwick, A. F. Mistr, Jr., and E. B. Croasdale, Reactive
Management: A Key to Survival in the 1990s, IEEE Transactions on
Power Systems, Vol. 10, No. 2, pp. 1036-1043, May 1995.
Using SCADA real-time data from BPAs system and data from other utilities via
Inter-utility Data Exchange as input, PSST creates a State Estimator case. This
identifies any overloads or voltage problems for the power system as it is running
at that time. The State Estimator runs automatically every 15 minutes. This
State Estimator case can be then used as input to a Powerflow program, when
the user can study the system further by adjusting generation levels, increasing
or decreasing Intertie schedules or creating additional outages.
At this time, BPA is involved in an intensive updating of the data base and a
thorough evaluation of program performance and the anticipated completion date
of this evaluation is March 1, 1997. The system was installed using a 1989
WSCC database, and while BPA Operations has kept up to date with changes
on the BPA power system, it is necessary to update the models of the
interconnected utilities. BPA currently has a inter-utility data exchange with 6
utilities (BC Hydro, Puget Sound Power & Light, PG&E, Portland General Electric,
Seattle City Light, and PacifiCorp) and expects to have data exchange with 5
additional utilities by summer of 1997. The system model must be improved to
include the additional detail for those utilities. As BPA receives more real-time
measurements from interconnected utilities, the accuracy of the results improve
and confidence in predicted power system performance is increased.
ESCA is developing a new program that will calculate the total transmission
capability (TTC) of transmission paths for current system conditions, and look
ahead 30 days using a load and outage forecast to predict future transmission
path capacity. BPA is involved in the development of this product because this
program will enhance power system security by predicting conditions where
maximum capability on the transmission paths is reduced.
COMPUTING INDICES
An index is computed for each of the following areas (voltage control areas):
VSA index computation was weighted values of power plants/SVC reactive power
outputs and reactive capabilities. Thirty power plants or SVCs are currently
specified. See attached spreadsheet.
Weights are based on relative sensitivity and plant size. We try to avoid current
limiting at the most sensitive plants, but recognize that current limiting at a small
plant is not critical. A large plant close to the voltage control area will have a high
weight. The individual weights will be entries in matrix of voltage security areas
and power plants/SVCs.
Wi = DQi x Qmxi
SDQi SQmxi
where DQi is plant reactive power increases for an increase in area load. The
first term ranges from 0 to 1 and a large values means that, for a load increase,
generation response at plant i is large compared to other plants; this is the
sensitivity term. The second term also ranges from 0 to 1, and a large value
means that the plant reactive power capability is large.
Different power flow modeling assumptions will also affect sensitivities to some
degree. Its probably better, for example, to neglect high side voltage control.
Sensitivities can be computed from a base case by repeat power flow simulation.
Additional reactive power load can be applied at area busses, with the power
plant reactive power changes observed. Alternatively, percentage change in a
power flow load in the zone or zones of the voltage control area can be made,
with plant reactive power changes observed. Its also possible to compute the
linear sensitivities from the Jacobian matrix of a solved power flow case.
Description: On-Line VSA will continuously monitor the voltage stability of the
grid using state estimated data from the energy management system (EMS). On-
line VSA will complete an assessment of voltage security of the current system
state within 20 minutes for large systems, providing various security indices for
the operator and a list of contingencies that could lead to instability. It will also
identify measures to mitigate an evolving voltage collapse, such as load
curtailment, re-dispatching and emergency loading of VAR sources, etc. The tool
will be capable of performing on-line studies of various operating decision
strategies, such as options for power transfers and scheduling planned outages
for maintenance.
Benefits: For voltage stability constrained utilities, VSA will allow increased
utilization of grid assets while maintaining the reliability of power delivery. This is
achieved through the calculation of limits dynamically using actual system
conditions. A typical benefit might be a 15% increase in transmission capacity
across a constrained 1,000 MW interface. Assuming 30% of this increased
capacity could be used on average over a year and a $10/Mwhr price differential
across the interface, this could result in an annual financial benefit of $3.9 million.
* Six months after the new BC Hydro EMS becomes fully operational.
Page 1 of 6
1.0 GENERAL
2.0 RESPONSIBILITIES
The System Control Centre (SCC) will direct Vancouver Island Control Centre
(VIC) and Lower Mainland Control Centre (LMC) to arm and disarm the automatic
undervoltage load shedding scheme as required in accordance with system
conditions. VIC and LMC will be responsible to ensure connected load is
available for shedding at all times and to allocate loads for shedding in different
blocks. After load shedding has occurred, load restoration shall be coordinated
by SCC.
Each subsystem monitors three key station bus voltages and a designated group
of units for its VAR reserve, which is the remaining VAR boost capacity of the
group. If the bus voltage drops below a set level or if the VAR reserve drops
below a set level, its sensor will key a continuous signal to either VIC or LMC. At
the control centre, the VAR reserve is ANDED with three different combinations
of two bus voltages. After time delay (Ts2), load shedding will start and continue
in incremental blocks until the initiating conditions have reset. A reduced time
delay (Ts1) is used if both subsystems have simultaneously initiated load
shedding. Refer to Attachment 1 of this order for a functional logic diagram of the
scheme.
Dynamic VAR source: total VAR output of VIT S/C1, S/C2, S/C3 and
S/C4 (in % of its "true" capacity).
Notes: 1) "True" capacity is based on connected units with AVRs in service. Any
unit with its AVR out of service do not contribute to "true" capacity.
3) If the positive VAR capacity of a unit is restricted below its normal rating,
P&C personnel will be notified to re-program new rating in the PLC.
3.3 Settings
The automatic undervoltage load shedding scheme will be armed and disarmed
by VIC and LMC under direction from SCC. The scheme is not required and
should not be armed at times when transmission lines are taken OOS for voltage
control during light load condition.
c) when the total dynamic VAR reserve for VIT S/Cs is less
than 135 MVAR;
d) when the total dynamic VAR reserve for BUT S/Cs is less
than 360 MVAR.
The SCC Dispatcher will notify the VIC and LMC Dispatcher of any
impending requirements to initiate manual load shedding based on
system conditions described in Section 4.0 of this order. This is the first
alert.
The SCC Dispatcher will cancel the first alert when load shedding is no
longer required.
While in the first alert and after occurrence of a contingency that will
impact the ability of the electric system to maintain voltage stability (e.g.
loss of a major 500 kV transmission line or a heavily loaded lower
voltage transmission line, or loss of major voltage support equipment),
the SCC Dispatcher will immediately inform the VIC and/or LMC
Dispatcher of the outage. The SCC Dispatcher will also delegate
responsibility to the ACC Dispatcher to monitor the key VAR reserve
group and substation voltages in the area and initiate manual load
shedding if these monitored quantities remain above or below their
settings described in section 3.3.
When communication between SCC and the ACC is not immediate but
the ACC Dispatcher can see that a major disturbance has occurred and
the VAR reserve and voltages in the control area are steadily declining,
the ACC Dispatcher can unilaterally take operating responsibility on an
emergency basis to shed load.
2. If the VAR output on the 2. If the VAR output from the BUT
VIT S/C units drop below 90% plant drops below 65% of its
of its capacity, no further load capacity, no further load
shedding is required. shedding is required.
The amount of load to shed in the first block must be greater than 125 MW.
Indicates Revisions
The closing control logic is designed to ensure that C231 will be automatically switched
on during significant loss of Idahos Eastern resources during high import conditions from the
Northwest. The additional flows superimposed on the 230kV lines into Boise Bench from the
West, during these circumstances, could lead to a drain in our reactive resources if adequate
voltage support is not available in our area. It is mainly for this reason, that timely switching of
the new bank (C231) is critical to prevent further voltage decline. The proposed scheme is
intended to achieve just that.
The actual control logic should preferably be implemented via a Programmable Logic
Controller, although both a relay with a flexible enough programmable logic or the current EMS
system might be capable of the task. Use of a PLC will provide ease of implementation and the
higher degree of flexibility needed to accommodate future modifications that may be required
under different system conditions. Additionally a simpler hardware/software interface and
broader accessibility to field personnel will cut down, when needed, the effective re-
programming time to a fraction of that obtained through the other two alternatives.
As indicated in the appended schematic, automatic closure of the bank can take place via
three separate criteria. The first and more sensitive one is through a power flow level detector
(PWR), which will normally be enabled, when the total MW flows into Boise Bench, on the
Brownlee lines is above a pre-determined minimum. This supervisory condition should reduce
unnecessary closing operations of the bank. With power flows above the prescribed minimum,
meeting either one of the following conditions will suffice for bank closure:
a) DPWR : MW change into Boise Bench on either one of the Brownlee lines above a
prescribed minimum. Note that each line is to be monitored separately and the change
measured over an adjustable time interval.
b) D I : Current magnitude change on either one of the Brownlee lines above a prescribed
minimum. Highest loaded phase or average phase loading is an acceptable compromise. Note
that each line is to be monitored separately and the change measured over an adjustable time
interval.
c) Qcond : Reactive or Current magnitude on either one of the Boise Bench synchronous
condensers over a prescribed minimum. Again if current measurement is to be used, either
highest loaded phase or average phase loading will suffice. Condition should be present over
an adjustable time interval.
The second and third criteria, are independent of transfer level information and are
intended to pick up on the more gradual changes in line loading and/or serve as last resort
indicators of the need for capacitor switching. One is based on line overload (based on highest
loaded phase, or average phase current on either one of the Brownlee lines) exceeding a pre-
determined limit for an adjustable time period. The last criterion relies on the 230kV and the
Here the intend is to utilize the intelligence already built into the automatic closing
control logic, to securely and dependably initiate the undervoltage load shedding program. It
should be noted that the proposed schemes function is to provide the go or no-go decision to the
load shedding program currently executed by the EMS system. Further improvements, in the
reliability of the scheme will require redundant communication and a more secure tripping logic
at the remote sites (load dropping stations).
If the capacitor closing signal is asserted (86-AUX, also with adjustable delay on drop
out) and there is no current being drawn by the shunt bank after a pre-determined time delay
(T1), the load shedding program will be initiated. Similarly if an undervoltage condition (UV2),
also based on 230 and 138kV bus voltages continues to exist for an adjustable time period
permission is given to initiate the load shedding program. At this time the EMS will initiate
automatic tripping of up to 300 MW under heavy load conditions. Additionally the dispatcher has
access to a manual load shedding page within the EMS, that could be use to expedite shedding
additional 500 MW of load.
43SW
PWR
IO/L UV1
DPWR DI QCOND
UNDERVOLTAGE LOAD SHEDDING IO/L : Current magnitude level detector on either one of
the Brownlee lines.
PROGRAM INITIALIZER. (2 sec time delay or sample time).
86-AUX T1 UV2
In a letter to area utilities, dated March 26,1997, the Sacramento Valley Study
Group (SVSG) recommended that area utilities implement an under voltage (UV)
load shedding program. The following UV load shedding parameters were
recommended:
SMUD 230 MW
PG&E 150 MW
City of Roseville 20 MW
Total 400 MW
2. This scenario was modified (by raising Southern New Mexico Import (SNM)
level) in order to bring the V-Q curve instability point as close to zero
MVAR margin point as possible. This case was then stored (instability
case). The instability case was used in order to determine the potentially
unstable voltage levels at every bus in EPE system.
3. The ALIS version of the instability case was used in a single contingency
analysis; each transmission line in EPE service area was dropped and
voltages at every distribution substation were monitored.
4. The results of the V-Q curve and single contingency analysis were
combined with the following information available for each substation: the
size of load, the number of distribution feeders, the location of the
substation relative to the generation and SVG, the availability of
supervisory control at this substation, the type of customers this substation
serves (restricted buses), the nature of load at this substation (power
factor), the physical limitation for the installation of the load shedding relay
equipment. The following is the list of factors derived from the V-Q curve
and contingency analysis:
The relative impact of MVAR injection at the considered bus on the 345 kV
system. (The voltage impact of MVAR injection at the considered bus was
monitored at all EPE 345 kV busses and plotted on the V-Q curve for the
considered bus. This impact can be visualized from the V-Q curve as the
slope of the voltage for the monitored busses (345 kV busses) relative to
the slop of the voltage of the considered bus.
5. The trigger voltage for each bus were determined as the lowest voltage
between following voltages.
6. The time delay was set for each bus in every category. The time delay was
spread in the range of 5-11 seconds in order to avoid false trips due to the
transient conditions as well as to protect the system from an excessive and
unnecessary load tripping (if all relays would operate simultaneously).
The time delay for load tripping following activation of the trigger
voltage is set for each bus (the time delay used varies in between
five and eleven seconds, with one second increment).
PSE, SCL, SPUD, TCL, BPA use undervoltage relays to trip breakers supplying
buses, feeders, or subtransmission lines. The breakers and their loads are
grouped into three levels to mitigate increasing severity of reduced voltage. At
each level, sufficient breakers are selected so that each utility trips about 5% of
their total load. The sum of loads at all three levels in the Puget Sound Basin is
about 1800 MW, which is 15% of a total 12,000 MW winter peak.
(a) Each utility trips breakers serving 5% of load when monitored bus
voltages fall to 90% or lower of normal for a minimum of 3.5 seconds.
(b) Each utility trips breakers serving 5% of load when monitored bus
voltages fall to 92% or lower of normal for a minimum of 5.0 seconds.
(c) Each utility trips breakers serving 5% of load when monitored bus
voltages fall to 92% or lower of normal for a minimum of 8.0 seconds.
Before Schultz Substation was constructed, BPA applied direct load tripping to an
aluminum smelter for a double line loss of Coulee-Raver 500 kV. Schultz
Substation reduced the need for load tripping by connecting the lines with
breakers in the middle, along with two other 500 kV lines. The direct load tripping
is not employed at this time.
Introduction
The consensus view of the panel was an expression of acceptance of the above
600/500 MVAR V-Q margin when substantiated by P-V analysis or another
complementary method. V-Q analysis alone is not convincing. It was brought out
that the demonstrated MW margin for the P-V cases presented is in line with
industry practices. In the near term, the panel endorses a 600/400= MVAR
margin for both planning and operating purposes. This margin may be relaxed at
a later date as better tools and information become available which reduce the
elements of uncertainty. Other methods of analysis, particularly dynamic
simulation, should be applied in marginal cases to confirm in detail that the
Councils Criteria are met. The above recommendations apply to disturbances
both inside and outside the NW system as they relate to the COI.
Other VAR related issues which should be addressed to improve system VAR
capability for reliability purposes include:
At the December 5-6, 1996 meeting BPA presented study results based on a
post-contingency 600 MVAR V-Q margin for N-1 events and a 400 MVAR V-Q
margin for N-2 common mode events. The Panel had two general concerns.
First, it was important to the Panel that the work be benchmarked to indicate what
MW margin this corresponds to; the panel suggested use of P-V analysis.
Studies presented by BPA at the January 27-28 meeting using P-V analysis
helped confirm the results of the V-Q analysis. A list of uncertainties that should
be given consideration in establishing planning and operating margin includes:
The consensus view of the panel was an expression of acceptance of the above
600/500 MVAR V-Q margin when substantiated by P-V analysis or another
complementary method. V-Q analysis alone is not convincing. It was brought out
that the demonstrated MW margin for the P-V cases presented is in line with
industry practices. In the near term, the panel endorses a 600/400= MVAR
margin for both planning and operating purposes. This margin may be relaxed at
a later date as better tools and information become available which reduce the
elements of uncertainty. Other methods of analysis, particularly dynamic
simulation, should be applied in marginal cases to confirm in detail that the
Councils Criteria are met. The above recommendations apply to disturbances
both inside and outside the NW system as they relate to the COI.
Other Comments
The Panel emphasized the need for P-V analysis to verify the rating of the
Interties and also to establish voltage limits by which the dispatcher can be
warned.
Until the interconnected systems develop a VAR exchange policy and controls
are in place to adhere to such policies, planners should make their VAR
margins sufficient to handle the likely VAR exchanges indicated by
experience.
Key among various System Reactive Study assumptions was the reactive margin
against which system performance should be evaluated and used to establish the
COI and PDCI maximum transfer levels based on Northwest system
performance. Such reactive margins are not established by the Councils
Reliability Criteria.
The Panel endorses a post-contingency V-Q analysis margin of not less than 600
MVAR at critical 500-kV buses for N-1 outages and 400 MVAR for N-2 common
mode outages. The adequacy of these margins was supported by P-V analysis.
These margins should be used in the near term for both planning and operating
purposes. (See Appendix C).
The Panel also recommends that these margins be periodically reviewed. The
margins may be relaxed at a later date as better tools and information become
available which reduce the elements of uncertainty. Other methods of analysis,
particularly dynamics simulation, should be applied in marginal cases to confirm
in detail that the Councils Reliability Criteria are met. It is recognized that system
changes and operating experience must also be factored into any proposed
future reactive margin increases or decreases.
Introduction
Sufficient reactive resources must be located throughout the electric systems, with a balance
between static and dynamic characteristics. Both static and dynamic reactive power resources are
needed to supply the reactive power requirements of customer demands and the reactive power
losses in the transmission and distribution systems, and provide adequate system voltage support
and control. They are also necessary to avoid voltage instability and widespread system collapse
in the event of certain contingencies. Transmission systems cannot perform their intended
functions without an adequate reactive power supply.
Dynamic reactive power support and voltage control are essential during power system
disturbances. Synchronous generators, synchronous condensers, and static var compensators
(SVCs and STATCOMs) can provide dynamic support. Transmission line charging and series
and shunt capacitors are also sources of reactive support, but are static sources.
Reactive power sources must be distributed throughout the electric systems among the
generation, transmission, and distribution facilities, as well as at some customer locations.
Because customer reactive demands and facility loadings are constantly changing, coordination
of distribution and transmission reactive power is required. Unlike active or real power (Mws),
reactive power (MVAR) cannot be transmitted over long distances and must be supplied locally.
Standard
S1. Reactive power resources, with a balance between static and dynamic
characteristics, shall be planned and distributed throughout the interconnected
transmission systems to ensure system performance as defined in Categories A, B,
and C of Table I in the I.A. Standards on Transmission Systems.
Measurements
M2. Generation owners and transmission providers shall work jointly to optimize
the use of generator reactive power capability. These joint efforts shall
include:
Guides
G1. Transmission owners should plan and design their reactive power facilities so
as to ensure adequate reactive power reserves in the form of dynamic reserves
at synchronous generators, synchronous condensers, and static var
compensators (SVCs and STATCOMs) in anticipation of system disturbances.
For example, fixed and mechanically-switched shunt compensation should be
used to the extent practical so as to ensure reactive power dynamic reserves at
generators and SVCs to minimize the impact of system disturbances.
G3. At continuous rated power output, new synchronous generators should have an
overexcited power factor capability, measured at the generator terminals, of
0.9 or less and an underexcited power factor capability of 0.95 or less. If a
synchronous generator does not meet this requirement, the generation owner
should make alternate arrangements for supplying an equivalent dynamic
reactive power capability to meet the areas reactive power requirements.
G4. Reactive power compensation should be close to the area of high reactive
power consumption or production.
G7. Power flow simulation of contingencies, including P-V and V-Q curve
analyses, should be used and verified by dynamic simulation when steady-
state analyses indicate possible insufficient voltage stability margins.
Introduction
Electric systems that experience heavy loadings on transmission facilities with limited reactive
power control can be vulnerable to voltage instability. Such instability can cause tripping of
generators and transmission facilities resulting in loss of customer demand as well as system
collapse. Since voltage collapse can occur suddenly, there may not be sufficient time for operator
actions to stabilize the systems. Therefore, a load shedding scheme that is automatically activated
as a result of undervoltage conditions in portions of a system can be an effective means to
stabilize the interconnected systems and mitigate the effects of a voltage collapse.
It is imperative that undervoltage relays be coordinated with other system protection and control
devices used to interrupt electric supply to customers.
Standards
S2. All UVLS programs shall be coordinated with generation control and
protection systems, underfrequency load shedding programs, Regional
load restoration programs, and transmission protection and control
programs.
Measurements
M2. Those entities owning or operating UVLS programs shall ensure that their
programs are consistent with any Regional UVLS programs and that exist
including automatically shedding load in the amounts and at locations,
voltages, rates, and times consistent with any Regional requirements. (S1)
M3. Each Region shall maintain and annually update an UVLS program database.
This database shall include sufficient information to model the UVLS program
in dynamic simulations of the interconnected transmission systems. (S1)
M4. Those entities owning or operating UVLS programs shall periodically (at least
every five years or as required by changes in system conditions) conduct and
document a technical assessment of the effectiveness of the design and
implementation of its UVLS program. Documentation of the UVLS technical
assessment shall be provided to the appropriate Regions and NERC on
request. (S1)
M5. Those entities owning or operating UVLS programs shall have a maintenance
program to test and calibrate their UVLS relays to ensure accuracy and reliable
operation. Documentation of the implementation of the maintenance program
shall be provided to the appropriate Regions and NERC on request. (S1)
M6. Those entities owning or operating an UVLS program shall analyze and
document all system undervoltage events below the initiating set points of
their UVLS programs. Documentation of the analysis shall be provided to the
appropriate Regions and NERC on request. (S1)
Guides
G1. UVLS programs should be coordinated with other system protection and
control programs (e.g., timing of line reclosing, tap changing, overexcitation
limiting, capacitor bank switching, and other automatic switching schemes).
G2. Automatic UVLS programs should be coordinated with manual load shedding
programs.
G3. Manual load shedding programs should not include, to the extent possible,
customer demand that is part of an automatic UVLS program.
REFERENCES
19. "Power System Stability and Control", Book published by EPRI, Chapter 14,
P. Kundur, 1994.
21. System Protection and Control, NERC Planning Standards and Guides,
NERC Engineering Committee, Interconnection Dynamics Working Group.
22. Criteria and Countermeasures for Voltage Collapse, CIGR Task Force
38.02.17, CIGR Brochure No. 101, October 1995.
26. Standard Load Models for Power Flow and Dynamic Performance
Simulation, IEEE Task Force on Load Representation for Dynamic
Performance, IEEE Transactions on Power Systems, Vol. 10, No. 3,
August 1995.
28. BPA Blue Ribbon Panel Report, Bonneville Power Administration, May
1997.